Polling has a strong allure in an election for many reasons. It appears to make it possible to call the “horse race” before the steeds come through the home stretch. It also has the appearance of objectivity, since it is supposedly based on a dark science that few people know enough to question. Polls are a story that falls into the laps of reporters, allowing them to write a piece on an election without any more work than reading a column of numbers.
Yet polls are nowhere near adequate for describing an election, even without getting into the difficulties of the math. Like so many things in our world today, polling fails because of a flawed assumption at the heart of it long before the technical stuff that jazzes it into what appears to be irrefutable facts. It’s not simply a matter of who is included in the poll, either – it’s the simple fact that democracy belongs to those who show up, not a percentage of a population.
I like to use the Minnesota Governor’s race as an example of the problem because it is a very dynamic situation. Our election this 2 November will be the fourth one in a row confounded by a strong third-party presence in the center. Yet that isn’t the real complication.
Any poll that does not have an appropriate sample is going to be skewed. There are many ways commonly discussed as problems in this regard, including the ability to reach cell phone only voters. That can be solved by what we call “cross tabs”, or a glimpse at the population polled to see if it matches who showed up at the last election – by age, race, gender, or whatever you like. John Zogby became famous for adjusting his polls if his sample appeared skewed, goosing the results to get what he thought he would have if the sample was better.
The problem at the heart of any poll is that people don’t necessarily show up on election day if they aren’t “into it”. A poll which does not make a strong attempt to identify “likely voters” needs to be ignored right out of the box for this reason, but there’s even more to it than that. Elections aren’t decided by percentages, they are decided by actual numbers of votes (as we all learned in our last Senate election). Take a look at the last five Governor’s races here in Minnesota, leaving out the Reform / Independence Party results:
There is a strong Republican base that has grown steadily along with the population, standing now at about 1.1 million voters. Other than 1998, the year Jesse Ventura was elected, they have been quite solid. We can reasonably expect that this year Tom Emmer, the Republican, will get that same vote total unless he screws up badly.
The DFL (Democratic Farmer Labor Party, what we call Democrats) doesn’t have that reliable base. Voters can peel off to the Independence Party and vote for Tom Horner, but they also might not show up. That’s been an issue dating back to before there was a third party in the center and it continues today. When referred to at all, it’s the “enthusiasm gap” that Democrats face after the excitement of 2008.
If you are running a DFL campaign for Governor your mission is a simple one: turn out more than 1.1 million people for your candidate. Period. Some of them will come from “turning” Independence Party voters your way, but a lot come from what’s called “Get Out the Vote” (GOTV). When people are excited and wouldn’t miss the election for the world, they will show up. Most of those marginal voters are Democrats and they get you over the top of the Republicans who are there pretty much no matter what.
That is what is in front of Mark Dayton right now. His campaign has to get that many people to show up at the polls. For this reason, a GOTV strategy has to be at the center of any successful DFL campaign, which is to say that your people have to be identified, energized, and ready to not just vote but help get other people to the polls any way they can.
There are implications for this reality far beyond polling. The typical “left/right” arguments don’t swing elections much at all unless they are useful in energizing marginal voters – and most of the arguments I see are far more likely to turn potential voters off to politics in general. This also explains the standard Republican strategy, the opposite of GOTV, where things get so nasty that a lot of people throw up their hands and hate all “politics” as they know it, losing interest in elections completely.
But for polling the message is very clear – an election is won by votes, not percentages. Most reporters make a point of staying away from campaigning because of their allegiance to “objectivity”, but that only means that they have trouble understanding the perspective that a campaign manager has to have – every vote counts. The result is a belief in polls and percentages that far exceeds their actual value.
Don’t be fooled by it. Democracy is always a matter of who shows up. Please do so this year and make your vote count. Thank you.